CN111368168A - Big data-based electricity price obtaining and predicting method, system and computer-readable storage medium - Google Patents

Big data-based electricity price obtaining and predicting method, system and computer-readable storage medium Download PDF

Info

Publication number
CN111368168A
CN111368168A CN202010171840.XA CN202010171840A CN111368168A CN 111368168 A CN111368168 A CN 111368168A CN 202010171840 A CN202010171840 A CN 202010171840A CN 111368168 A CN111368168 A CN 111368168A
Authority
CN
China
Prior art keywords
electricity price
data
obtaining
value
big data
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Withdrawn
Application number
CN202010171840.XA
Other languages
Chinese (zh)
Inventor
刘留
张骁
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Foshan Aiwente Intelligent Technology Co ltd
Original Assignee
Foshan Aiwente Intelligent Technology Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Foshan Aiwente Intelligent Technology Co ltd filed Critical Foshan Aiwente Intelligent Technology Co ltd
Priority to CN202010171840.XA priority Critical patent/CN111368168A/en
Publication of CN111368168A publication Critical patent/CN111368168A/en
Withdrawn legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/951—Indexing; Web crawling techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/004—Artificial life, i.e. computing arrangements simulating life
    • G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/044—Recurrent networks, e.g. Hopfield networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/084—Backpropagation, e.g. using gradient descent
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/02—Marketing; Price estimation or determination; Fundraising
    • G06Q30/0201—Market modelling; Market analysis; Collecting market data
    • G06Q30/0206—Price or cost determination based on market factors
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Electricity, gas or water supply
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
    • Y02E40/00—Technologies for an efficient electrical power generation, transmission or distribution
    • Y02E40/70—Smart grids as climate change mitigation technology in the energy generation sector
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
    • Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
    • Y04S10/00—Systems supporting electrical power generation, transmission or distribution
    • Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
    • Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
    • Y04S50/00—Market activities related to the operation of systems integrating technologies related to power network operation or related to communication or information technologies
    • Y04S50/14—Marketing, i.e. market research and analysis, surveying, promotions, advertising, buyer profiling, customer management or rewards

Abstract

The invention provides a big data-based electricity price obtaining and predicting method, a system and a computer-readable storage medium, wherein the method comprises the following steps: constructing a power price obtaining model according to the sample data; acquiring text data related to current electricity prices through a big data information platform; performing text segmentation on the collected text data by adopting a word bank and a dictionary-free word segmentation algorithm, and extracting value information by adopting a feature extraction algorithm; and inputting the value information into an electricity price acquisition model to acquire the electricity price state of the current time period. The method and the system can accurately acquire the electricity price state at the current stage, are convenient for local power undertaking units to quickly, accurately and comprehensively acquire the latest electricity price policy information, digest the electricity price policy in time and improve the execution level of the electricity price. Meanwhile, the method and the device accurately predict the electricity price state in the future period by establishing an electricity price prediction model, are beneficial to power generation enterprises to establish the optimal power generation strategy and quotation strategy, and are beneficial to users to reduce the use cost according to the electricity price and the demand.

Description

Big data-based electricity price obtaining and predicting method, system and computer-readable storage medium
Technical Field
The invention relates to the field of electric power, in particular to a big data-based electricity price obtaining and predicting method, system and computer-readable storage medium.
Background
Electric power energy is an important material basis for economic growth and social development, and electric power belongs to one of important resources controlled by the state. Compared with the common commodity price, the electricity price carrying function is more complex, and the electricity price carrying device has stronger political requirements besides common economic and social requirements. Therefore, the electricity price policy is often used as an important means for the national macro-regulation and control of energy strategies and is also a compass for guiding the healthy development of the power industry.
Currently, in the key period of deepening electric power system innovation, the frequency of new price policies issued by electric power supervision departments is gradually accelerated, and as a bearing unit of local electric power supply, the latest price policy information needs to be rapidly, accurately and comprehensively acquired, the price policy needs to be digested in time, the trend of the policy is reasonably pre-judged, policy risks are effectively avoided, and the execution level of the price is improved.
Meanwhile, in the electric power market, the electricity price prediction also has important significance. For power generation enterprises, the method is beneficial to making an optimal power generation strategy and an optimal quotation strategy, and is also beneficial to reducing the use cost of users according to the electricity price and the demand. Especially, under the condition of using the real-time electricity price, a user can adjust the own electricity utilization plan according to the predicted electricity price, so that the use cost can be reduced, and the load at the peak moment of electricity utilization can be transferred to the valley moment of the whole power grid, so that the electricity use conditions at different moments tend to be average, and the electricity generation cost is reduced.
Based on the above requirements, it is urgently needed to provide an accurate electricity price obtaining and predicting method based on big data.
Disclosure of Invention
In order to solve at least one technical problem described above, the present invention proposes a big data-based electricity price acquisition and prediction method, system, and computer-readable storage medium.
In order to achieve the above object, a first aspect of the present invention provides a big data-based electricity price obtaining and predicting method, including:
constructing a power price obtaining model according to the sample data;
acquiring text data related to current electricity prices through a big data information platform;
performing text segmentation on the collected text data by adopting a word bank and a dictionary-free word segmentation algorithm, and extracting value information by adopting a feature extraction algorithm;
and inputting the value information into an electricity price acquisition model to acquire the electricity price state of the current time period.
In this aspect, after obtaining the electricity price state of the current time period, the method further includes:
inputting the obtained electricity price state into a database;
carrying out normalization preprocessing on the historical electricity price data in the database;
training and constructing a power price prediction model according to the normalized and preprocessed historical power price data;
and predicting the electricity price state of the preset node and/or the preset time period in the future according to the electricity price prediction model.
In this scheme, the calculation formula of the normalization preprocessing is as follows:
Figure BDA0002409430950000021
wherein x' represents the normalized preprocessed data; x is the number ofmaxMaximum value, x, representing historical electricity price data in a databaseminRepresents the minimum value of the historical electricity rate data in the database.
In the scheme, a power price prediction model is trained and constructed according to the historical power price data after normalization preprocessing, and the method specifically comprises the following steps:
determining a BP neural network structure;
initializing the weight and the threshold of the BP neural network and the position and the speed of particles in the particle swarm;
inputting electricity price sample data for training, and calculating the fitness value of the particles;
searching individual extreme values Gbest of the particles, judging whether the current adaptability value is superior to the extreme value Gbest or not, and if so, updating the extreme value Gbest of the particles;
searching a group extreme value Zbest, judging whether the current adaptive values of all the particles are superior to the extreme value Zbest, and if so, updating the extreme value Zbest of the particles;
according to the formula
Figure BDA0002409430950000031
Updating the velocity V of the particles and according to the formula
Figure BDA0002409430950000032
Updating the position X of the particle, wherein c1Is the self-acceleration coefficient of the particle, c2Is a global acceleration factor, mu1，μ2Is distributed in [0,1 ]]Random number of interval, rho is constraint factor;
according to the formula
Figure BDA0002409430950000033
Updating the weight omega; wherein, TmaxIs the number of iterations, ωmax，ωminThe maximum and minimum weight is obtained, and t is the current iteration value;
checking whether a termination condition is met, if so, stopping iteration to obtain the optimal weight and threshold of the BP neural network;
and calculating errors, updating the weight and the threshold, checking whether an ending condition is met, stopping iteration if the ending condition is met, and outputting the optimal weight and the threshold of the BP neural network.
In this embodiment, after checking whether the termination condition is satisfied, the method further includes:
and if the termination condition is not met, recalculating the fitness value of the particle and entering the next iteration.
In this scheme, after determining whether the current fitness value is better than the extremum Gbest or whether the current fitness value of all the particles is better than the extremum Zbest, the method further includes:
if not, according to the formula
Figure BDA0002409430950000034
Updating the velocity V of the particles and according to the formula
Figure BDA0002409430950000035
The position X of the particle is updated.
In the scheme, the BP neural network uses a Sigmoid equation as a transfer function between each neuron, and the relation between input x and output f (x) of the neuron is as follows:
Figure BDA0002409430950000036
in this embodiment, after performing normalization preprocessing on the historical electricity price data in the electricity price database, the method further includes:
respectively substituting the lowest temperature, the highest temperature and the average temperature of the same type of days in the database into the membership function of low temperature, medium temperature and high temperature;
according to the maximum membership principle, fuzzy values of the lowest temperature, the highest temperature and the average temperature of the same type of days are respectively calculated;
and training the electricity price prediction model by taking fuzzy values of the lowest temperature, the highest temperature and the average temperature of the same type of days as input quantities.
The second aspect of the present invention also provides an electricity price acquiring and predicting system, including: a memory and a processor, the memory including a big data based electricity price obtaining and predicting method program, the electricity price obtaining and predicting protecting method program when executed by the processor implementing the steps of a big data based electricity price obtaining and predicting method as described above.
The third aspect of the present invention also provides a computer-readable storage medium, which includes a big data-based electricity price obtaining and predicting method program, and when the big data-based electricity price obtaining and predicting method program is executed by a processor, the steps of the big data-based electricity price obtaining and predicting method are implemented.
According to the method, the text data related to the electricity price policy is obtained through the big data information, the feature extraction is carried out through a machine learning method, the valuable information is mined, and then the electricity price state of the current stage is accurately obtained based on the valuable information. Meanwhile, the method accurately predicts the electricity price state in the future period by establishing an electricity price prediction model, is beneficial to power generation enterprises to establish an optimal power generation strategy and a quotation strategy, and is also beneficial to users to reduce the use cost according to the electricity price and the demand.
Additional aspects and advantages of the invention will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the invention.
Drawings
FIG. 1 shows a flow chart of a method of electricity price acquisition of the present invention;
FIG. 2 is a flow chart illustrating a method of electricity price prediction in accordance with the present invention;
FIG. 3 illustrates a flow chart of a method of training a power price prediction model of the present invention;
FIG. 4 illustrates a flow chart of a method of the present invention for training a power price prediction model using temperature data;
FIG. 5 illustrates a block diagram of an electricity price acquisition and prediction system of the present invention.
Detailed Description
In order that the above objects, features and advantages of the present invention can be more clearly understood, a more particular description of the invention will be rendered by reference to the appended drawings. It should be noted that the embodiments and features of the embodiments of the present application may be combined with each other without conflict.
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention, however, the present invention may be practiced in other ways than those specifically described herein, and therefore the scope of the present invention is not limited by the specific embodiments disclosed below.
Fig. 1 shows a flowchart of an electricity rate obtaining method of the present invention.
As shown in fig. 1, a first aspect of the present invention provides a big data-based electricity price obtaining and predicting method, including:
s102, constructing a power price obtaining model according to sample data;
s104, acquiring text data related to current electricity prices through a big data information platform;
s106, performing text word segmentation on the collected text data by adopting a word bank and a word segmentation algorithm without a dictionary, and extracting value information by adopting a feature extraction algorithm;
and S108, inputting the value information into an electricity price obtaining model to obtain the electricity price state of the current time period.
It should be noted that the technical solution of the present invention can be operated in a terminal device such as a PC, a mobile phone, a PAD, and the like.
The text data may be a text related to an electricity price policy, which belongs to public data, is published on government web pages such as development and modification agencies and price offices, and is published news at professional vertical portals such as the power industry. The sample data is history text related to an electricity price policy, and the electricity price obtaining model is obtained by machine learning based on the history text.
It should be noted that, in step S104, the text content may be crawled from various websites (e.g., Baidu, Google, etc.) by using a web crawler. Web crawlers can be roughly classified into the following types according to system structure and implementation technology: general Purpose Web crawlers (General Purpose Web Crawler), Focused Web Crawler (Focused Web Crawler), Incremental Web Crawler (Incremental Web Crawler), Deep Web Crawler (Deep Web Crawler). The present invention preferably employs application-focused web crawlers and incremental web crawlers.
The focused web crawler refers to a web crawler that selectively crawls pages related to a predefined topic. Compared with the general web crawler, the focusing crawler only needs to crawl pages related to the theme, so that hardware and network resources are greatly saved, and the requirements of specific groups on information in specific fields can be well met. The incremental web crawler is a crawler which updates downloaded web pages in an incremental mode and only crawls newly generated or changed web pages, and can ensure that the crawled web pages are as new as possible to a certain extent. Compared with the web crawler which crawls and refreshes the pages periodically, the incremental crawler only crawls the newly generated or updated pages when needed, does not download the unchanged pages again, can effectively reduce the data download amount, timely updates the crawled pages, and reduces the consumption in time and space.
It should be noted that, in step S106, the word segmentation algorithm based on the lexicon includes a forward maximum matching, a forward minimum matching, a reverse matching, a word-by-word traversal matching method, and the like. The algorithm has the characteristics of easy realization and simple design, but the correctness of word segmentation depends on the built word stock to a great extent; the word segmentation algorithm based on the dictionary is based on the statistics of word frequency, the occurrence frequency of any two adjacent words in the original text is taken as a word for statistics, the higher the occurrence frequency is, the higher the possibility of becoming a word is, and when the frequency exceeds a certain preset threshold value, the word is taken as a word for indexing. The method can effectively extract unknown words. The invention combines the two word segmentation algorithms to effectively improve the precision of word segmentation.
Feature extraction refers to extracting metadata about text, including descriptive features (such as name, date, size, type, etc.) and semantic features (such as author, organization, title, content, etc.). For descriptive characteristics with fixed formats, such as dates, types and the like, simple regular expressions can be used for extraction, and for semantic characteristics, firstly, characteristic representation is required, wherein certain characteristic items (such as terms or descriptions) are used for representing texts, and the characteristic items are processed, so that unstructured text processing is realized. The feature representation model includes boolean logic type, Vector Space Model (VSM), probabilistic type, hybrid type, and the like. After the features are represented, a rating function is constructed based on a feature extraction algorithm, each feature is evaluated, the features are ranked according to scores, and a preset number of features with the highest scores are selected. The merit function includes: information Gain (Information Gain), Expected cross entropy (Expected cross entropy), Mutual Information (Mutual Information), and text Evidence Weight (The Weight of Evidence for text). But is not limited thereto.
Fig. 2 shows a flow chart of a power rate prediction method of the present invention.
As shown in fig. 2, after acquiring the power rate status of the current period, the method further includes:
s202, inputting the acquired electricity price state into a database;
s204, carrying out normalization preprocessing on the historical electricity price data in the database;
s206, training and constructing a power price prediction model according to the normalized and preprocessed historical power price data;
and S208, predicting the electricity price state of the preset node and/or the preset time period in the future according to the electricity price prediction model.
It should be noted that, the calculation formula of the normalization preprocessing is as follows:
Figure BDA0002409430950000081
wherein x' represents the normalized preprocessed data; x is the number ofmaxMaximum value, x, representing historical electricity price data in a databaseminRepresents the minimum value of the historical electricity rate data in the database.
It should be noted that the database is used for storing historical electricity price states at various time stages and other data influencing the electricity price states; the power price prediction model is a BP neural network model based on a particle swarm algorithm.
FIG. 3 illustrates a flow chart of a method of training a power price prediction model of the present invention.
As shown in fig. 3, training and constructing a power price prediction model according to the normalized preprocessed historical power price data specifically includes:
step 1, determining a BP neural network structure;
it should be noted that the BP neural network is a multi-layer feedforward neural network that adjusts the network weights by using a back propagation algorithm. In the network training, errors between an output layer and actual output samples are reversely transmitted to an input layer through hidden layers, and the errors of neurons in all layers are calculated layer by layer to correct all connection weights and thresholds, so that ideal errors are finally achieved.
Preferably, the invention adopts a 3-layer BP neural network structure, takes a Sigmoid equation as a transfer function between each neuron, and the relation between the input x and the output f (x) of the neuron is as follows:
Figure BDA0002409430950000082
step 2, initializing the weight and the threshold of the BP neural network and the position and the speed of particles in the particle swarm;
it should be noted that the weight of the BP neural network may be initialized according to the following calculation formula:
Figure BDA0002409430950000083
Vjk＝Vjk+μXjek；
wherein i is 1,2, …, n; j ═ 1,2, …, l; k is 1,2, …, m; vij、VjkIs the BP neural network connection weight, p (i), XjInput layer input and hidden layer output respectively; μ is the learning rate, ekIs an error;
it should be noted that the threshold of the BP neural network may be initialized according to the following calculation formula:
Figure BDA0002409430950000091
bk＝bk+ek；
wherein j is 1,2, …, l; k is 1,2, …, m; a isj、bkFor BP neural network hidden layer and output layer valvesThe value is obtained.
Step 3, inputting electricity price sample data for training, and calculating the fitness value of the particles;
it should be noted that, the historical electricity price data in the database can be taken as the electricity price sample data for training; the calculation formula of the fitness value F is as follows:
Figure BDA0002409430950000092
wherein N is the number of training samples; c is the number of the output network neurons;
Figure BDA0002409430950000093
the electricity price ideal output value of the j network output node of the i sample is obtained;
Figure BDA0002409430950000094
and outputting the actual output value of the electricity price of the j network output node for the i sample.
Step 4, searching individual extreme values Gbest of the particles, judging whether the current adaptability value is superior to the extreme values Gbest, and if so, updating the extreme values Gbest of the particles; if not, according to the formula
Figure BDA0002409430950000095
Figure BDA0002409430950000096
Updating the velocity V of the particles and according to the formula
Figure BDA0002409430950000097
The position X of the particle is updated.
Step 5, searching a group extreme value Zbest, judging whether the current adaptive values of all the particles are superior to the extreme value Zbest, and if so, updating the extreme value Zbest of the particles; if not, according to the formula
Figure BDA0002409430950000098
Figure BDA0002409430950000099
Updating the velocity V of the particles and according to the formula
Figure BDA00024094309500000910
The position X of the particle is updated.
Step 6, according to the formula
Figure BDA00024094309500000911
Updating the velocity V of the particles and according to the formula
Figure BDA00024094309500000912
Updating the position X of the particle, wherein c1Is the self-acceleration coefficient of the particle, c2Is a global acceleration factor, mu1，μ2Is distributed in [0,1 ]]The random number of the interval, ρ, is a constraint factor.
Step 7, according to the formula
Figure BDA0002409430950000101
Updating the weight omega; wherein, TmaxIs the number of iterations, ωmax，ωminAnd t is the current iteration value, and is the maximum and minimum weight.
Step 8, checking whether a termination condition is met, if so, stopping iteration to obtain the optimal weight and threshold of the BP neural network; and if the termination condition is not met, recalculating the fitness value of the particle and entering the next iteration.
Step 9, calculating errors, updating weights and thresholds, checking whether a finishing condition is met, if so, stopping iteration, and outputting the optimal weights and thresholds of the BP neural network; if not, error is recalculated.
Furthermore, after the electricity price result is preliminarily predicted by using the BP neural network, the state transition rule of the prediction error can be determined according to the Markov chain, and then the electricity price prediction result is corrected.
The Markov chain calculation formula in the Markov chain prediction model is as follows: xn＝X0PnWherein X isnAnd X0The state probability vectors at n time and the initial time, respectively, and P is a markov state transition probability matrix.
Inputting the electricity price sample data into the BP neural network to obtain a preliminary prediction result C of the prediction time npAnd calculating the absolute percentage error, and dividing the Markov state interval by adopting a fuzzy C-mean clustering method to obtain m states Q1,Q2,Q3,…,QmAnd counting the absolute percentage error according to the divided state space to obtain a Markov state transition table, and solving a Markov state transition probability matrix according to the Markov state transition table by using the state transition table as follows:
Figure BDA0002409430950000111
Figure BDA0002409430950000112
wherein, PijIs in a state QiTransfer to state Q through 1 stepjTransition probability of, NjFor Markov state transition table QjNumber of samples of (1), NijIs in a state QiTransfer to state Q through 1 stepjThe frequency of (2).
Calculating formula X from the Markov chain described aboven＝X0PnCalculating to obtain a state probability vector X at time nn，XnMiddle maximum probability value XpCorresponding state QpThe absolute percentage error state of the predicted time n is obtained; preliminary prediction result C of the BP neural networkpAnd correcting, wherein the corrected prediction result is as follows:
C*＝Cp(1+δ*)；
Figure BDA0002409430950000113
wherein, deltaa，δbIs in a state QpUpper and lower bounds of the interval (2).
FIG. 4 illustrates a flow chart of a method of the present invention for training a power price prediction model using temperature data.
As shown in fig. 4, after the normalization preprocessing of the historical power rate data in the power rate database, the method further includes:
s402, substituting the lowest temperature, the highest temperature and the average temperature of the same type of days in the database into low-temperature, medium-temperature and high-temperature membership function respectively;
s404, respectively solving fuzzy values of the lowest temperature, the highest temperature and the average temperature of the same type of days according to the maximum membership degree principle;
and S406, training a power price prediction model by using fuzzy values of the lowest temperature, the highest temperature and the average temperature of the same type of day as input quantities.
When the electricity price prediction model is trained using the temperature data as an input amount, the temperature data may be subjected to a blurring process.
The low-temperature membership function adopts a small trapezoidal distribution as follows:
Figure BDA0002409430950000121
the intermediate temperature membership function adopts the following formula in a middle trapezoidal distribution:
Figure BDA0002409430950000122
the membership function of high temperature adopts a large trapezoidal distribution as follows:
Figure BDA0002409430950000123
the lowest temperature T of the same type of daylBy substituting the above three formulas, the membership { T } of three states can be obtainedl1，Tl2，Tl3}. According to the principle of maximum membership, Tl＝max{Tl1,Tl2,Tl3}, known as TlThe fuzzy set to which it belongs. For maximum temperature ThAnd the average temperature TaThe same method can be used to find the three statesDegree of membership and its value.
Furthermore, holidays are also important factors influencing the fluctuation of the electricity price, and the electricity price state of special holidays can be calculated by establishing a holiday model; the specific calculation method comprises the following steps:
step 1, predicting the special daily electricity price PSpecially for treating diabetesIn the mean time, the average value P of the power price of the weekend in one month before a special day is calculatedSpecial front 0Then, the average value P of the electricity prices of the working days of one week before the special day is calculatedSpecial front 1。
Step 2, find the average value P 'of the weekend electricity prices within this particular previous month of the previous year'Special front 0Average value P 'of working days of one week before this day'Special front 1。
Step 3, solving the increase proportional relation of the electricity price, wherein the calculation formula is as follows:
Figure BDA0002409430950000131
Figure BDA0002409430950000132
solving the historical data of the special daily electricity prices of the previous two years according to the principle of importance, lightness and weakness, wherein the proportion of the two years is qa1，qa2，qb1，qb2The daily electricity prices of the first two years are respectively PSpecial 1，PSpecial 2(ii) a Calculating and obtaining the weighted average proportional relation q according to the following calculation formula1，q2，
Figure BDA0002409430950000135
Is a weighted average of the holiday electricity prices of the first two years,
q1＝α·qa1+(1-α)·qa2；
q2＝α·qb1+(1-α)·qb2；
Figure BDA0002409430950000133
preferably, α is 0.73 and β is 0.11, but not limited thereto.
Step 4, calculating the special day electricity price, taking different weights according to the similarity of the weekend electricity price and the special day electricity price; the calculation formula is as follows:
Figure BDA0002409430950000134
FIG. 5 illustrates a block diagram of an electricity price acquisition and prediction system of the present invention.
As shown in fig. 5, the second aspect of the present invention also proposes an electricity price acquiring and predicting system 5, the electricity price acquiring and predicting system 5 including: a memory 51 and a processor 52, wherein the memory 51 includes a program of a big data-based electricity price obtaining and predicting method, and when the program of the electricity price obtaining and predicting protecting method is executed by the processor, the following steps are implemented:
constructing a power price obtaining model according to the sample data;
acquiring text data related to current electricity prices through a big data information platform;
performing text segmentation on the collected text data by adopting a word bank and a dictionary-free word segmentation algorithm, and extracting value information by adopting a feature extraction algorithm;
and inputting the value information into an electricity price acquisition model to acquire the electricity price state of the current time period.
It should be noted that the system of the present invention can be operated in a terminal device such as a PC, a mobile phone, a PAD, etc.
It should be noted that the Processor may be a Central Processing Unit (CPU), other general-purpose Processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), an off-the-shelf Programmable Gate Array (FPGA) or other Programmable logic device, a discrete Gate or transistor logic device, a discrete hardware component, and so on. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
Further, after acquiring the power rate state of the current period, the method further includes:
inputting the obtained electricity price state into a database;
carrying out normalization preprocessing on the historical electricity price data in the database;
training and constructing a power price prediction model according to the normalized and preprocessed historical power price data;
and predicting the electricity price state of the preset node and/or the preset time period in the future according to the electricity price prediction model.
Further, the calculation formula of the normalization preprocessing is as follows:
Figure BDA0002409430950000141
wherein x' represents the normalized preprocessed data; x is the number ofmaxMaximum value, x, representing historical electricity price data in a databaseminRepresents the minimum value of the historical electricity rate data in the database.
Further, training and constructing a power price prediction model according to the normalized and preprocessed historical power price data, which specifically comprises the following steps:
determining a BP neural network structure;
initializing the weight and the threshold of the BP neural network and the position and the speed of particles in the particle swarm;
inputting electricity price sample data for training, and calculating the fitness value of the particles;
searching individual extreme values Gbest of the particles, judging whether the current adaptability value is superior to the extreme value Gbest or not, and if so, updating the extreme value Gbest of the particles;
searching a group extreme value Zbest, judging whether the current adaptive values of all the particles are superior to the extreme value Zbest, and if so, updating the extreme value Zbest of the particles;
according to the formula
Figure BDA0002409430950000151
Updating the velocity V of the particles and according to the formula
Figure BDA0002409430950000152
Updating the position X of the particle, wherein c1Is the self-acceleration coefficient of the particle, c2Is a global acceleration factor, mu1，μ2Is distributed in [0,1 ]]Random number of interval, rho is constraint factor;
according to the formula
Figure BDA0002409430950000153
Updating the weight omega; wherein, TmaxIs the number of iterations, ωmax，ωminThe maximum and minimum weight is obtained, and t is the current iteration value;
checking whether a termination condition is met, if so, stopping iteration to obtain the optimal weight and threshold of the BP neural network;
and calculating errors, updating the weight and the threshold, checking whether an ending condition is met, stopping iteration if the ending condition is met, and outputting the optimal weight and the threshold of the BP neural network.
Further, after checking whether the termination condition is satisfied, the method further includes:
and if the termination condition is not met, recalculating the fitness value of the particle and entering the next iteration.
Further, after determining whether the current adaptive value is better than the extreme value Gbest or whether the current adaptive values of all the particles are better than the extreme value Zbest, the method further includes:
if not, according to the formula
Figure BDA0002409430950000154
Updating the velocity V of the particles and according to the formula
Figure BDA0002409430950000155
The position X of the particle is updated.
Further, the BP neural network uses Sigmoid equation as a transfer function between each neuron, and the relation between neuron input x and output f (x) is:
Figure BDA0002409430950000156
further, after the normalization preprocessing is performed on the historical electricity price data in the electricity price database, the method further includes:
respectively substituting the lowest temperature, the highest temperature and the average temperature of the same type of days in the database into the membership function of low temperature, medium temperature and high temperature;
according to the maximum membership principle, fuzzy values of the lowest temperature, the highest temperature and the average temperature of the same type of days are respectively calculated;
and training the electricity price prediction model by taking fuzzy values of the lowest temperature, the highest temperature and the average temperature of the same type of days as input quantities.
The third aspect of the present invention also provides a computer-readable storage medium, which includes a big data-based electricity price obtaining and predicting method program, and when the big data-based electricity price obtaining and predicting method program is executed by a processor, the steps of the big data-based electricity price obtaining and predicting method are implemented.
According to the method, the text data related to the electricity price policy is obtained through the big data information, the feature extraction is carried out through a machine learning method, the valuable information is mined, and then the electricity price state of the current stage is accurately obtained based on the valuable information. Meanwhile, the method accurately predicts the electricity price state in the future period by establishing an electricity price prediction model, is beneficial to power generation enterprises to establish an optimal power generation strategy and a quotation strategy, and is also beneficial to users to reduce the use cost according to the electricity price and the demand.
In the several embodiments provided in the present application, it should be understood that the disclosed apparatus and method may be implemented in other ways. The above-described device embodiments are merely illustrative, for example, the division of the unit is only a logical functional division, and there may be other division ways in actual implementation, such as: multiple units or components may be combined, or may be integrated into another system, or some features may be omitted, or not implemented. In addition, the coupling, direct coupling or communication connection between the components shown or discussed may be through some interfaces, and the indirect coupling or communication connection between the devices or units may be electrical, mechanical or other forms.
The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units; can be located in one place or distributed on a plurality of network units; some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
In addition, all the functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may be separately regarded as one unit, or two or more units may be integrated into one unit; the integrated unit can be realized in a form of hardware, or in a form of hardware plus a software functional unit.
Those of ordinary skill in the art will understand that: all or part of the steps for realizing the method embodiments can be completed by hardware related to program instructions, the program can be stored in a computer readable storage medium, and the program executes the steps comprising the method embodiments when executed; and the aforementioned storage medium includes: a mobile storage device, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
Alternatively, the integrated unit of the present invention may be stored in a computer-readable storage medium if it is implemented in the form of a software functional module and sold or used as a separate product. Based on such understanding, the technical solutions of the embodiments of the present invention may be essentially implemented or a part contributing to the prior art may be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the methods described in the embodiments of the present invention. And the aforementioned storage medium includes: a removable storage device, a ROM, a RAM, a magnetic or optical disk, or various other media that can store program code.
The above description is only for the specific embodiments of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art can easily conceive of the changes or substitutions within the technical scope of the present invention, and all the changes or substitutions should be covered within the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the appended claims.

Claims (10)

1. A big data based electricity price acquisition and prediction method, the method comprising:
constructing a power price obtaining model according to the sample data;
acquiring text data related to current electricity prices through a big data information platform;
performing text segmentation on the collected text data by adopting a word bank and a dictionary-free word segmentation algorithm, and extracting value information by adopting a feature extraction algorithm;
and inputting the value information into an electricity price acquisition model to acquire the electricity price state of the current time period.
2. The big data-based electricity price obtaining and predicting method according to claim 1, wherein after obtaining the electricity price status of the current period, the method further comprises:
inputting the obtained electricity price state into a database;
carrying out normalization preprocessing on the historical electricity price data in the database;
training and constructing a power price prediction model according to the normalized and preprocessed historical power price data;
and predicting the electricity price state of the preset node and/or the preset time period in the future according to the electricity price prediction model.
3. The big data-based electricity price obtaining and predicting method according to claim 2, wherein the calculation formula of the normalization preprocessing is as follows:
Figure FDA0002409430940000011
wherein x' represents the normalized preprocessed data; x is the number ofmaxMaximum value, x, representing historical electricity price data in a databaseminRepresents the minimum value of the historical electricity rate data in the database.
4. The big data-based electricity price obtaining and predicting method according to claim 2, wherein the training of the electricity price prediction model according to the normalized preprocessed historical electricity price data specifically comprises:
determining a BP neural network structure;
initializing the weight and the threshold of the BP neural network and the position and the speed of particles in the particle swarm;
inputting electricity price sample data for training, and calculating the fitness value of the particles;
searching individual extreme values Gbest of the particles, judging whether the current adaptability value is superior to the extreme value Gbest or not, and if so, updating the extreme value Gbest of the particles;
searching a group extreme value Zbest, judging whether the current adaptive values of all the particles are superior to the extreme value Zbest, and if so, updating the extreme value Zbest of the particles;
according to the formula
Figure FDA0002409430940000021
Updating the velocity V of the particles and according to the formula
Figure FDA0002409430940000022
Updating the position X of the particle, wherein c1Is the self-acceleration coefficient of the particle, c2Is a global acceleration factor, mu1，μ2Is distributed in [0,1 ]]Random number of interval, rho is constraint factor;
according to the formula
Figure FDA0002409430940000023
Updating the weight omega; wherein, TmaxIs the number of iterations, ωmax，ωminIn order to be the maximum of the minimum weight,t is the current iteration value;
checking whether a termination condition is met, if so, stopping iteration to obtain the optimal weight and threshold of the BP neural network;
and calculating errors, updating the weight and the threshold, checking whether an ending condition is met, stopping iteration if the ending condition is met, and outputting the optimal weight and the threshold of the BP neural network.
5. The big data-based electricity price obtaining and predicting method according to claim 4, wherein after checking whether a termination condition is satisfied, the method further comprises:
and if the termination condition is not met, recalculating the fitness value of the particle and entering the next iteration.
6. The big-data-based electricity price obtaining and predicting method according to claim 4, wherein after determining whether the current adaptation value is better than the extreme value Gbest or whether all the current adaptation values of the particles are better than the extreme value Zbest, the method further comprises:
if not, according to the formula
Figure FDA0002409430940000024
Updating the velocity V of the particles and according to the formula
Figure FDA0002409430940000025
The position X of the particle is updated.
7. The big data-based electricity price obtaining and predicting method according to claim 4, wherein the BP neural network uses Sigmoid equation as transfer function between each neuron, and the relation between neuron input x and output f (x) is as follows:
Figure FDA0002409430940000031
8. the big data-based electricity price obtaining and predicting method according to claim 2, wherein after the normalization preprocessing of the historical electricity price data in the electricity price database, the method further comprises:
respectively substituting the lowest temperature, the highest temperature and the average temperature of the same type of days in the database into the membership function of low temperature, medium temperature and high temperature;
according to the maximum membership principle, fuzzy values of the lowest temperature, the highest temperature and the average temperature of the same type of days are respectively calculated;
and training the electricity price prediction model by taking fuzzy values of the lowest temperature, the highest temperature and the average temperature of the same type of days as input quantities.
9. An electricity price acquisition and prediction system characterized by comprising: a memory and a processor, the memory including a big data-based electricity price obtaining and predicting method program, the electricity price obtaining and predicting protecting method program when being executed by the processor implementing the steps of a big data-based electricity price obtaining and predicting method according to any one of claims 1 to 8.
10. A computer-readable storage medium, characterized in that the computer-readable storage medium includes a big-data-based electricity price obtaining and predicting method program, which when executed by a processor, implements the steps of a big-data-based electricity price obtaining and predicting method according to any one of claims 1 to 8.
CN202010171840.XA 2020-03-12 2020-03-12 Big data-based electricity price obtaining and predicting method, system and computer-readable storage medium Withdrawn CN111368168A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010171840.XA CN111368168A (en) 2020-03-12 2020-03-12 Big data-based electricity price obtaining and predicting method, system and computer-readable storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010171840.XA CN111368168A (en) 2020-03-12 2020-03-12 Big data-based electricity price obtaining and predicting method, system and computer-readable storage medium

Publications (1)

Publication Number Publication Date
CN111368168A true CN111368168A (en) 2020-07-03

Family

ID=71211833

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010171840.XA Withdrawn CN111368168A (en) 2020-03-12 2020-03-12 Big data-based electricity price obtaining and predicting method, system and computer-readable storage medium

Country Status (1)

Country Link
CN (1) CN111368168A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112232886A (en) * 2020-10-30 2021-01-15 南方电网能源发展研究院有限责任公司 Electricity price probability prediction method, system, computer equipment and storage medium
CN116362421A (en) * 2023-05-31 2023-06-30 天津市普迅电力信息技术有限公司 Energy supply distribution prediction system and method based on comprehensive overall analysis of energy sources

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112232886A (en) * 2020-10-30 2021-01-15 南方电网能源发展研究院有限责任公司 Electricity price probability prediction method, system, computer equipment and storage medium
CN116362421A (en) * 2023-05-31 2023-06-30 天津市普迅电力信息技术有限公司 Energy supply distribution prediction system and method based on comprehensive overall analysis of energy sources
CN116362421B (en) * 2023-05-31 2023-10-10 天津市普迅电力信息技术有限公司 Energy supply distribution prediction system and method based on comprehensive overall analysis of energy sources

Similar Documents

Publication Publication Date Title
Shen et al. Wind speed prediction of unmanned sailboat based on CNN and LSTM hybrid neural network
Zhang et al. A deep learning approach for daily tourist flow forecasting with consumer search data
Kirelli et al. Sentiment analysis of shared tweets on global warming on twitter with data mining methods: a case study on Turkish language
US20230222409A1 (en) Apparatus for Determining Role Fitness While Eliminating Unwanted Bias
CN114519145A (en) Sequence recommendation method for mining long-term and short-term interests of users based on graph neural network
CN111368168A (en) Big data-based electricity price obtaining and predicting method, system and computer-readable storage medium
CN112380344A (en) Text classification method, topic generation method, device, equipment and medium
Liu et al. EACP: An effective automatic channel pruning for neural networks
CN115080868A (en) Product pushing method, product pushing device, computer equipment, storage medium and program product
Yuan et al. Deep learning from a statistical perspective
Dai et al. Research on mathematical model of box office forecast through BP neural network and big data technology
Guo et al. Prediction of index trend based on lstm model for extracting image similarity feature
Zhao et al. A hierarchical attention recommender system based on cross-domain social networks
Song et al. A novel text classification approach based on Word2vec and TextRank keyword extraction
Wenjie et al. A NOVEL MODEL FOR STOCK CLOSING PRICE PREDICTION USING CNN-ATTENTION-GRU-ATTENTION.
Xie et al. Ternary sentiment classification of airline passengers’ twitter text based on bert
Li et al. Adaptive trading system based on LSTM neural network
CN115270988A (en) Fine adjustment method, device and application of knowledge representation decoupling classification model
Koromyslova et al. Feature selection for natural language call routing based on self-adaptive genetic algorithm
CN114881172A (en) Software vulnerability automatic classification method based on weighted word vector and neural network
Dong et al. A hierarchical network with user memory matrix for long sequence recommendation
Lai et al. Chinese stock market prediction based on multifeature fusion and TextCNN
CN111242519A (en) User characteristic data generation method and device and electronic equipment
Chenyang et al. Research on Knowledge Classification Based on KNN and Naive Bayesian Algorithms
CN109284379B (en) Adaptive microblog topic tracking method based on two-way quantity model

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
WW01 Invention patent application withdrawn after publication
WW01 Invention patent application withdrawn after publication

Application publication date: 20200703